NOTES
26446
chain-branching rate expression. Similarly, if deuterium is substituted for hydrogen, one would expect the rates of reactions e and g to be decreased much more than that of reaction f. Hence one expects the deuterium pressure to appear in the chain branching rate expression a t lower temperatures than those at which the hydrogen pressure appears. Using the data of Frost and Alyea7for deuterium-oxygen explosions and a value of 0.5 for KI’/KI of equation (4)the lines of Figure 5 for DZwere obtained. It is apparent that the substitution of Dz for Hz practically obliterates the temperature effect observed for Hz.
The authors wish to express their sincere appreciation for many helpful criticisms made by Dr. L. S. Kassel.
Summary An equation has been suggested for representing the explosion limits of hydrogen-oxygen mixtures as a function of the partial pressures of the reactants. This equation is shown to be qualitatively consistent with a series of possible reactions for the chain branching and chain breaking mechanisms; and it is in quantitative agreement with most of the available data. PITTSBURGH, PA.
(7) Frost and Alyea, THISJOURNAL, 66, 1251 (1934).
Vol. 56
RECEIVED OCTOBER20, 1934
NOTES Boron Arsenate
As, 50.34. Found: B, 7.21, 7.27; As, 50.17,
BY WALTERC. SCHUMB AND WINSLOW H. HARTFORD 50.07, 50.20.
In connection with our study of the compounds of boron described elsewhere, we have prepared the relatively little known boron arsenate, and have determined certain of its properties. The method of Meyer3 for the preparation of boron phosphate was employed, but modifications were necessary due to the greater tendency of boron arsenate to hydrolyze. The best results were obtained by the following procedure: 100 g. of arsenic acid was dissolved in water, the solution was filtered from any suspended matter, heated to boiling, and 20 g. of boric acid then was added. The liquid was maintained at the boiling point until, suddenly, a viscous, white precipitate formed, which was washed rapidly by decantation, filtered and dried; 24.5 g. of a pure white product was obtained. This product still contained a small amount of boric acid, but when boiled with ethyl alcohol, in which boric acid is somewhat soluble and hydrolysis does not occur, a substance was obtained which analysis showed to be very nearly pure boron arsenate. Boron and arsenic were determined on separate samples, using the standard Chapin method for boron, and the U. S. Department of Agriculture method for arsenic.* Anal. Calcd. for BAs04: B, 7.27; (1) Schumb and Hartford, TEIS JOURNAL, 66, 2613 (1934). (2) Berger, Corn);. rend., 170,1492 (1920); Schulze, Nalurdss., 01, 582 (1933); 2. pbysik. Cbcm., EM, 216 (1934~;Gruner, 2. anorg. allgem. Chum., 019, 181 (1934). (3) hleyer, Be?., 90, 2919 (18891. (4) Chapin, THISJOURNAL, 30, 1687 (1908); U.S. Dept. of Agriculture, Bull. 107. p. 239.
The ready hydrolysis of boron arsenate is indicated by the facts that in contact with water the substance was wholly decomposed after three days, and that even in acid solution hydrolysis slowly progressed. The specific gravity of this substance was found to be, at room temperature, 3.40, a fact which, taken in comparison with the much lower specific gravities of the reactants, indicates a highly condensed internal structure. Its index of refraction (nD)was 1.66. It is readily decomposed by hot water or warm solutions of sodium carbonate or sodium hydroxide ; concentrated hydrochloric acid dissolves it on warming, forming arsenic trichloride and liberating chlorine. It is soluble in dilute sulfuric acid but not in the cold, concentrated acid. It is slowly dissolved by dilute or concentrated nitric acid. At red heat it is slightly decomposed without melting. PUBLICATION N O . 44 FROM THE RESEARCH LABORATORY OF INORGANIC CHEMISTRY MASSACEiUSETTS INSTITUTE OF TECHNOLOGY CAMBRIDGE, MASS. RECEIVED JANUARY 16, 1934
Occurrence and Position of the Principal Inflection Point in Certain Acid-Base Titration
Curves BY E. D. EASTMAN
This note comprises a refutation of the statements of P. S. Roller1 concerning my previous (1) Roller, TEIS JOURNAL, 64, a485 (1932).
--
Dec., 1934 work2 on this subject. It also includes certain new results and simplifications, and a recapitulation of the main theoretical expectations relating to the titration of monobasic acids by monacid bases. The passages (numbered for convenience, and in order of importance) from Roller’s paper with which I disagree are cited below, accompanied in each case by my comment. I (a) (Ref. 1, p. 3494).-‘‘The conclusion of Eastman that an inflection point may not appear in the (symmetrical) titration of a weak base is in the light of these results seen to be incorrect.” (b) (Ref. 1, pp. 3494-3495) .-“The deduction of Eastman that for the symmetrical titration of a weak acid by a weak base there is a titer deviation of the inflection point from the stoichiometric point is seen to be erroneous.” To make clear the differences in results implied in these statements, the behavior indicated by Equations 10 and 8 of my paper for the titration of a weak acid by an equally weak base will be given. At concentrations of either 1 M or 0.001 111,if K A = KB = lov5the principal inflection occurs when H + concentration is about 2% greater and the “titer” 0.0401, less than at the end-point. If KA = KB = lo-@, the inflection point, at either concentration, is still found3but deviates by approximately 30% in (H+) and 4.5% in titer from the end-point. If KA = K B = 10-7, the inflection no longer appears. According to Roller the inflection always appears if KA = K B , and always coincides with the end-point in this case. I have rechecked the derivation of my equations and calculations from them, confirming the above results. I have also checked these results from the expression for dpH/d V (pH = Sorensen number, ‘G’ = moles of base added) and find maxima at the positions found from Eq. 10 for and none the inflections with K = low6and if K = lo-’. I have also verified these results by direct substitution in Eq. 7 of (H+) values at intervals of 0.05 unit in PH, with the same results. I therefore r e a f f i the former conclusions. Qualitatively, the predicted disappearance of the inflection as the acid and base are progressively weakened is the normally expected result. On the other hand, the dev%tion of the inflection, when it does appear, from the end-point is disturbing in view of the apparent “symmetry” to (2) Eastman, (a)THIS JOURNAL, 47,332 (1925); (b) 60,418 (1928). (3) In the former paper the limit in K A for 0.001 M WBS incorrectly given as 1 0 - 6 .
NOTES
2647
which Roller refers. Actually, however, this symmetry does not exist. The ordinary plot of pH against titer is unsymmetrical about the endpoint (or any other point). The principle of symmetry cannot therefore be invoked in the present discussion. I1 (Ref. 1, p. 3495).-(a) Eastman’s “estimate of the titer deviation in the unsymmetrical titration by a strong base of a 1 molar weak acid with KA = 10-l2 is about ten-fold too low; (b) and his conclusion that the titer deviation for this titration does not vary with concentration of weak acid between 1 and mole/liter is plainly in error.” The calculation mentioned in (a) corresponded to K A = lo-” and not as assumed by Roller. It is correct for the former value. My conclusion mentioned in I1 (b) carried the proviso that the acid strength should be just sufficient (only integral powers of ten being considered in KA) to produce an inflection. With this proviso the conclusion is correct. Thus with concentration 1 M , the limiting strength of acid is KA = 10-l2 and the deviation in titer is about 3%. A t 0.001 M the limiting strength is KA = and the titer deviation again about 3%. These values are in complete agreement with Roller’s calculations. In the former paper the limiting acid strength was mistakenly taken as lo-” and for the respective concentrations, the titer deviation then corresponding to 0.3%. There was no intention to imply that for a fixed value of KA the titer deviation is independent of concentration. 111 (Ref. 1,p. 3486).-(a) Eastman’s “numerical result for the condition of the appearance of an inflection point was shown to be erroneous; (b) his other numerical results are also in error as shown below.” The errors mentioned in (a) are trivial and have been acknowledged previously (Ref. 2b). They resulted from the use of intervals (one power of 10) in (Hf) which were too large for solving the equation by successive approximations. Three of the calculations were affected by this error (one of these being recognized for the first time above). All other numerical results I believe to be correct under the (idealized) conditions to which they apply. Although this completes my reply to Roller’s criticisms, I shall take this opportunity of giving a more convenient form of Eq. 10 of my former
paper, with a few further results based upon it. The equation of condition for appearance of an inflection may be written, with the same symbols as previously except that H replaces (H+) as follows: (1
- Kw/KBH)(cKA/(KA+ HI2 4- Rw/HB + 1)
+ +
-
(Kw/KBH~)(cKA/(KA H) Kw/H - H) ~(RW/KBH 4-~)(cKAH/(KA-k H)* 4-Kw/H2) = 0
Using this equation for several cases not previously considered in which acid and base are both weak, I find that if KA or KB = the inflection occurs if the ionization constant of but the titrating base or acid is as large as not if it is = 10-lo. Similarly, if KA or KB is the inflection occurs if the ionizataken as tion constant of the titrating agent is but not if it is Indefinite elaboration of such calculations is possible. I believe, however, that all that is ordinarily required for correct orientation is provided in the following summary. If acid and base are both strong (ionization constant infinite) the inflection appears at all concentrations and always coincides with the endpoint. If the acid is weak and the added base strong the inflection appears if C K A> 2.7 X 10-13, as shown by Roller. The inflection precedes the end-point by about 3% in titer if cKA = 10-l2. This last value decreases approximately 10-fold for each 10-fold increase in cKA. With acid and base both weak, and with c between 0.001 M and 1 M , the numerical results above may be generalized by the approximation that the inflection will appear if KAKB 2 10-12, but not if KAKB 5 The inflection in these cases does not coincide with the end-point. The deviation in titer is nearly independent of concentration in the indicated range. It is of the order of 5% if K A K B = and again decreases about 10-fold for each 10-fold increase in KAKB. It will be observed that the generalization above limits the critical value of KAKB for the disappearance of the inflection only within one power of 10, and that it is based on examples in which neither constant exceeded It can no doubt be applied somewhat beyond this point, but must fail eventually if KA or KB is indefinitely increased. Further refinement of the criterion, within more restricted limits of K,is possible, but of little additional value. CHEMICAL LABORATORY RECEZVED AUGUST2, 1934 UNIVERSITY OF CALIFORNIA
BERKELEY, CALWORNIA
Vol. 56
NOTES
2648
An Investigation of the Benzoate Method for the Separation of Iron, Aluminum and Chromium and Some Suggested Changes for its Application to Qualitative Analysis BY LEOLEHRMAN AND
JACOB
KRAMER
Recently a method of separating aluminum, chromic and ferric ions, using ammonium benzoate as a precipitant, from the other ions of the third and fourth groups and magnesium was 0ffered.l While the procedure was devised essentially for quantitative work the authors say it can be used in a qualitative scheme. However, certain features of the method seem not to be clearly developed for the qualitative analysis of the common metals and others may not be readily adaptable. These points in question, which we thoroughly investigated are : (A) the application of the procedure for 500 mg. total metallic ion content, the usual amount for qualitative work, especially when it is mostly trivalent ; (B) diminution of the lengthy period of boiling required to precipitate the Cr+++; (C) the possible difficulty of filtering the precipitate representing 500 mg. of trivalent metallic ion; (D) reduction of the number of washings and possible change of the wash solution. Our investigation, the results of which are given below, showed that with certain changes the method can be used for the qualitative analysis of the common metals. With 500 mg. separately of each of the divalent ions of Groups 3 and 4 and Mg++, no precipitation is obtained with 10% ammonium benzoate reagent in acetic acid solution. For separate solutions of 500 mg. each of Al+++,C r + + +and Fe+++,90 ml., 60 ml. and 40 ml., respectively, of the precipitant are needed to cause complete precipitation after one minute of boiling. Mixtures of 250 mg. of each of the trivalent metallic ions, Fe+++and Cr+++,AI+++ and Cr+++,and Fe+++and AI+++, required 40 ml., 50 ml. and 60 ml., respectively, to give the same result. In all cases if not enough precipitant is originally added for complete precipitation, prolonged boiling has no effect. Thus it is possible to precipitate completely 500 mg. of Cr ++ + in a short period of boiling (one minute). The size of the precipitate obtained with 500 mg. of the trivalent metallic ions is rather large but can be contained in an 11 cm. filter paper if filtered with suction. Under these conditions two or three washings with 10 (1) Kolthoff, Stenger and Moskowitz, Tms (1934).
JOURNAL,
66, 812
POLYMERIZATION OF DIHYDROXYACETONE
Dec., 1934
ml. of 5% ammonium nitrate solution removes practically all of the divalent metallic ions. Finally 1 mg. of the divalent metallic ions a n be detected in the filtrate after precipitation of 500 mg. of any one of the trivalent or combination of them. Before testing the filtrate for the divalent
2649
ions the benzoic acid present is removed by evaporating until crystallization just begins, cooling and filtering with suction. D E P A R ~ E OF N TCHEMISTRY
mE
cm coLLBoeNEW
NEWYORE CIW
RECEIVED OCIOBER 13,1934
BIOLOGY. CARNBOIE rNSl"TUT10N OF WASHINGTON. AND T n E DlVlSION OF PLANT NUTRITION. UNIVERSITY OF CALIFORMA]
ICOM'RIBLJTION FROU THE D N I S I O N OF PLANT
Polymerization of Dihydroxyacetone BY H ~ O LHD STRAIN AND W. H. DORE x-Ray diffraction measurements demonstrate the polymerization of crystalline monomolecular dihydroxyacetone to dimolecular dihydroxyacetone. These measurements also indicate that the polymerization begins shortly after the preparation of the monomolecular form and that it is complete in twenty-five to thirty days. At the end of this period, higher polymers and condensation products were not detectable by x-ray, cryoscopic or solubility methods.
acetone were obtained a t intervals over a period of t h i y days. The results indicated that, a t room temperature and within the experimental error, polymerization to the dimolecular form was complete in twenty-five to thirty days and was independent of radiation with x-rays. The average molecular weight of the polymerized product determined by cryoscopic methods corresponds to that of dimolecular dihydroxyacetone.'
Experimental Dimolecular dihydroxyacetone, prepared by crystallization (from ethanol) of dihydroxyacetone which had stood for several months' and shown to be dimolecular by cryoscopic measurements of freshly prepared aqueous solutions, exhibited (when powdered) a clear and characteristic x-ray diffraction pattern (Column B, Table I). Monomolecular dihydroxyacetone,prepared by distillation of the dimolecular form%and shown to be monomolecular by cryoscopic measurements of freshly prepared aqueous solutions, likewise exFig. I.-x-Ray patterns of some forms of dihydroryhibited a clear and characteristic x-ray diffraction acetone: A, monomolecular dihydroxyacetone. rn. p. 7.2'; B, dmolecular dihydroxyacetone. m. p. 78-81'; pattern (Column A, Table I). After crystals of C, polymolecular dihydroxyacetone, m. p. 130-133"; the monomolecular dihydroxyacetone had stood D, polymolecular dihydroxyacetone. m. p. 230-260' for several days it was observed that the diffracwith dec.; E, polymolecular dihydroxyacetone, m. p. tion pattern resembled that obtained from the di230-250° with dec. molecular form even though molecular weight deAttempts were made t o determine whether or terminations demonstrated that the aystals were not higher polymers and condensation were composed of the monomolecular form. This indi(a) Shile q s t . 1 ~ of d i m o l w l a r dihydmxya&me ol iuadent cated that the monomolecular form was slowly optical, goniomeVie end h u e diillnnion m-urements polymerizing to the dimolecularform and, in order wlirem for ohtlined by cryde.Ilizin8 the sugar from ethanol. Optical to determine the rate of the polymerization, x-ray exemination revenled that the crystab were monoclinic but that they were twinned in some complex fashion nnd elear Lnuediffrnetion diffrsction patterns of monomolecular dihydroxy- patterns muld not be obtained. The results iadiented a unit cell
I
(1) Bertrand. CanPI. rmd.. 110. &(2,984 (1888); Ann. cEm. phyr.,
[sia. 216. 240 (IQM).
(2) Piwher pnd Mildbrnnd. B e . . 87, 707 (1924); Reeves 4 Renbom, B i a r k n ~J.,18.412 (1851).
m m W of t w o molcolln of dimolrvltir dihydroxyacetone. Single crystals of monomolecular dibydroxyamtoae muld not be obtained by crystdlizntion b e c a w of the rnoid m k m e r i u t i o n to tho dim0leEuIe.c form d u u i b e d in t h b paper.
